| 摘 要: 为适应问答系统智能化程度越来越高的特点，提出基于自然语言处理转化为SQL(Structured Query Language，结构化查询语言)语言技术的政务智能搜索引擎系统。用户通过输入自然语言问题直接获得相关数据，数据可以表格、图形等方式直观地显示。建立了融合SQL语法和增强列信息的算法模型SQL Model，利用NL2SQL(Natural Language Processing to Structured Query Language，自然语言转化为结构化查询语言)技术设计政务智能搜索引擎系统，并以某市的人口数据进行实验。实验结果表明，该技术可有效降低数据应用的复杂度，实现多维度复杂查询，降低业务部门数据搜索应用难度，提高政务数据搜索效率。 |
| 关键词: 问答系统；NL2SQL；政务智能搜索引擎系统；SQL Model |
|
中图分类号: TP31
文献标识码: A
|
| 基金项目: 深圳市基础研究重点项目(20200829114939001)；深圳市科技计划项目(GJHZ20180929154602092)；深圳市技术攻关面上项目(JCYJ20190809145407809). |
|
| Research on the Application of Government Affairs Intelligent Search Engine based on NL2SQL Technology |
|
YAO Junhua1, TANG Daijia2
|
( 1.Shenzhen Nanshan District Government Affairs Service Data Management Bureau, Shenzhen 518000, China ;
2.Shenzhen United Information Technology Company, Shenzhen 518000, China)
yaojunhuasz@163.com; tangdaijiasz@163.com
|
| Abstract: In order to accommodate the increasingly intelligent characteristics of the question answering system, this paper proposes an intelligent search engine system for government affairs based on the technology of transforming natural language processing into SQL (Structured Query Language) language. Users can directly obtain relevant data by inputting natural language questions, and the data can be displayed intuitively in tables, graphs, and so on. An algorithm model of SQL Model is established that incorporates SQL syntax and enhances column information. NL2SQL (Natural Language Processing to Structured Query Language) technology is used to design the intelligent search engine system for government affairs, and an experiment is carried out with the population data of a city. The experimental results show that this technology can effectively reduce the complexity of data application, realize multi-dimensional complex query, reduce the application difficulty of data search in departments, and improve the data search efficiency of government affairs. |
| Keywords: question answering system; NL2SQL; government affairs intelligent search engine system; SQL model |
